Abstract: In this study, we propose an approach to build fault-tolerant distributed real-time embedded systems. From a given system description and a given fault hypothesis, we generate automatically a fault-tolerant distributed schedule based on GSFR of the source algorithm onto the target architecture which minimizes the system’s run-time and tolerates buses communication failures. The scheduling algorithm proposed is dedicated to multi-bus heterogeneous architectures with multiple processors linked by several shared buses. It is based on passive redundancy and variable data fragmentation strategies which allow fast fault detection/retransmission and efficient use of buses, the size of each fragmented data depends on GSFR and the bus failure rates, variable fragment size allows reliable communication and maximize the reliability of the system. As this scheduling problem is NP-hard, we use a heuristic algorithm to obtain an approximate efficiently solution and we are able to show with simulation results that our approach can generally reduce the run-time overhead.